Certainly in 2020, Belgium counted “even suspected cases of COVID-19 in the national death toll, regardless of whether the deceased person was tested.”, while others counted only confirmed cases (can be very low if you don’t test much) or ignored deaths in nursing homes (https://www.dw.com/en/belgiums-coronavirus-overcounting-cont...)
USA's CDC estimate of the ratio is that the actual number of cases is 4.2*nbr of positive PCR tests. Which means that we only catch about 24% of the actual number of cases.
I would think also that countries like Italy which had big waves early have a more serious undercount (for total cases in the pandemic) as compared to countries which first got it somewhat later, when PCR tests were more generally available.
It's not just who had the most cases previously, though, because Belgium is very high again. Curious.